Description and Yield of Current Quality and Safety Review in Selected US Academic Emergency Departments.

OBJECTIVES: Quality and safety review for performance improvement is important for systems of care and is required for US academic emergency departments (EDs). Assessment of the impact of patient safety initiatives in the context of increasing burdens of quality measurement compels standardized, meaningful, high-yield approaches for performance review. Limited data describe how quality and safety reviews are currently conducted and how well they perform in detecting patient harm and areas for improvement. We hypothesized that decades-old approaches used in many academic EDs are inefficient and low yield for identifying patient harm. METHODS: We conducted a prospective observational study to evaluate the efficiency and yield of current quality review processes at five academic EDs for a 12-month period. Sites provided descriptions of their current practice and collected summary data on the number and severity of events identified in their reviews and the referral sources that led to their capture. Categories of common referral sources were established at the beginning of the study. Sites used the Institute for Healthcare Improvement's definition in defining an adverse event and a modified National Coordinating Council for Medication Error Reporting and Prevention (MERP) Index for grading severity of events. RESULTS: Participating sites had similar processes for quality review, including a two-level review process, monthly reviews and conferences, similar screening criteria, and a grading system for evaluating cases. In 60 months of data collection, we reviewed a total of 4735 cases and identified 381 events. This included 287 near-misses, errors/events (MERP A-I) and 94 adverse events (AEs) (MERP E-I). The overall AE rate (event rate with harm) was 1.99 (95% confidence interval = 1.62%-2.43%), ranging from 1.24% to 3.47% across sites. The overall rate of quality concerns (events without harm) was 6.06% (5.42%-6.78%), ranging from 2.96% to 10.95% across sites. Seventy-two-hour returns were the most frequent referral source used, accounting for 47% of the cases reviewed but with a yield of only 0.81% in identifying harm. Other referral sources similarly had very low yields. External referrals were the highest yield referral source, with 14.34% (10.64%-19.03%) identifying AEs. As a percentage of the 94 AEs identified, external referrals also accounted for 41.49% of cases. CONCLUSIONS: With an overall adverse event rate of 1.99%, commonly used referral sources seem to be low yield and inefficient for detecting patient harm. Approximately 6% of the cases identified by these criteria yielded a near miss or quality concern. New approaches to quality and safety review in the ED are needed to optimize their yield and efficiency for identifying harm and areas for improvement.

Observation and inpatient status: clinical impact of the 2-midnight rule.

BACKGROUND: In response to growing concern over frequency and duration of observation encounters, the Centers for Medicare and Medicaid Services enacted a rules change on October 1, 2013, classifying most hospital encounters of <2 midnights as observation, and those ≥2 midnights as inpatient. However, limited data exist to predict the impact of the new rule. OBJECTIVE: To answer the following: (1) Will the rule reduce observation encounter frequency? (2) Are short-stay (<2 midnights) inpatient encounters often misclassified observation encounters? (3) Do 2 midnights separate distinct clinical populations, making this rule logical? (4) Do nonclinical factors such as time of day of admission impact classification under the rule? DESIGN, SETTING AND PATIENTS: Retrospective descriptive study of all observation and inpatient encounters initiated between January 1, 2012 and February 28, 2013 at a Midwestern academic medical center. MEASUREMENTS: Demographics, insurance type, and characteristics of hospitalization were abstracted for each encounter. RESULTS: Of 36,193 encounters, 4,769 (13.2%) were observation. Applying the new rules predicted a net loss of 14.9% inpatient stays; for Medicare only, a loss of 7.4%. Less than 2-midnight inpatient and observation stays were different, sharing only 1 of 5 top International Classification of Diseases, 9th Revision (ICD-9) codes, but for encounters classified as observation, 4 of 5 top ICD-9 codes were the same across the length of stay. Observation encounters starting before 8:00 am less commonly spanned 2 midnights (13.6%) than later encounters (31.2%). CONCLUSIONS: The 2-midnight rule adds new challenges to observation and inpatient policy. These findings suggest a need for rules modification.